*2.4. Varicocele*

Varicocele is found in 5–10% of men with NOA [16]. Although several pathophysiological hypotheses have been proposed about the link between varicocele and NOA, no definite conclusions can be drawn [17]. Despite this, varicocele repair has been proposed to be beneficial in patients with NOA: a meta-analysis evaluating 16 studies for a total cohort of 344 azoospermic men who had undergone varicocele repair reported that 43.9% (151/344) of them had sperm in the ejaculate (sperm count was 1.82 ± 1.58 million/mL (95% CI: 0.98–2.77 millions/mL), sperm motility was 22.9% ± 15.5% (95% CI: 12.5–33.2%) 4.5 to 11 months after surgery; testicular biopsies were obtained in 8 out of 16 studies, histopathology demonstrating that the chance of having sperm in the ejaculate was significantly higher in patients with hypospermatogenesis (HS) compared to maturation arrest (MA) (OR: 2.35; 95% CI: 1.04–5.29; *p* = 0.04), and to Sertoli cell only syndrome (SCO) (OR: 12.0; 95% CI: 4.34–33.17; *p* < 0.001) [18]. However, since positive changes in the semen parameters following varicocele repair may not last forever, sperm cryopreservation is recommended [17]. The same meta-analysis reports the results of three studies evaluating the SRR in patients with varicocele, which was significantly greater in men with prior varicocele repair, compared to untreated patients (OR 2.65, 95% CI 1.69–4.14). Still, such studies were not devoid of selection bias. On the other hand, Schlegel and Kaufmann evaluated 138 patients with NOA and varicocele, 68 with a prior varicocelectomy, and 70 who did not undergo surgery: SRR was comparable in both groups (41/68 (60%) vs. 42/70 (60%), and did not vary with histopathological subcategories (26 vs. 38% in SCO, 53 vs. 47% in MA, and 96 vs. 96% in HS in patients with prior varicocelectomy or no treatment, respectively) [19]. Similarly, a study evaluating 860 patients with NOA, of whom 169 had prior history of varicocele repair, by means of a predictive model with varicocelectomy, age, prior sperm retrieval, testis volume, FSH, LH, testosterone level and diagnosis of KS as candidate predictors (all found to be predictive of SSR in univariate logistic regression), found that prior varicocelectomy was not predictive of SSR in multivariate logistic regression [20]. Given the conflicting results as above, well-designed randomized clinical trials are warranted to clarify whether varicocele repair may help in the managemen<sup>t</sup> of patients with NOA.

#### *2.5. Testis Volume*

Since seminiferous tubules contribute to approximately 80% of testis volume, this clinical parameter has been classically correlated with spermatogenesis. Indeed, a large sample size study (2.672 patients) demonstrated that testis volume correlates with sperm parameters and serum gonadotrophins levels [21], and men with testicular long axis 4.6 cm or less have been found to be more likely to have azoospermia, due to spermatogenic dysfunction [22]. Nevertheless, the correlation between testis volume and SSR in mTESE is not as intuitive as one would expect. On one hand, sperm may be retrieved even in patients with testis volume lower than 2 mL, with SRR being comparable to that of patients with larger testes (sample size = 1127 patients) [23]; on the other hand, patients with NOA due to early maturation arrest usually display normal testis volume, but have the worst chance of sperm retrieval [24].

Still, a meta-analysis evaluating 117 studies enrolling 21,404 patients found that testis volume significantly predicted SRR, specifically a mean volume higher than 12.5 mL predicted a SRR > 60%, with an accuracy of 86.2 ± 0.01% (*p* < 0.0001) and a specificity and sensitivity of 73% and 74% respectively; notably, the study design of the studies included in the analysis was heterogeneous with regard to patients' clinical characteristics and the surgical procedure applied (cTESE or mTESE) [9]. Indeed, a meta-analysis that included only studies evaluating patients with NOA who had undergone mTESE (5 studies with a total of 1764 cases) found that testis volume had limited value in predicting positive sperm retrieval in patients with NOA (AUC 0.63), mostly due to low specificity (sensitivity 80%, 95% CI: 0.78–0.83, specificity 35%, 95% CI: 0.32–0.39) [25]. It may be concluded, therefore, that patients with NOA with small testes should not be discouraged from attempting mTESE in the hand of skilled urologists.

**Table 1.** Comparison of sperm retrieval rates in patients with NOA with normal karyotype or Klinefelter syndrome.


hCG, human chorionic gonadotropin, nkNOA, patients with NOA with normal karyotype, KS, patients with Klinefelter syndrome, SRR, sperm retrieval rate, T, serum testosterone level.

#### **3. Hormonal Parameters**

Follicle-stimulating hormone (FSH) and testosterone (T) are both required to promote full spermatogenesis; in addition, their serum levels reflect both the pituitary and testicular function in physiological and pathological conditions. Indeed, the measurement of FSH and T serum levels represents the minimal initial hormonal evaluation of the azoospermic men, to distinguish between primary and secondary testicular failure [37]. Their role as predictors of spermatogenesis in patients with NOA is, however, questionable.

#### *3.1. Follicle-Stimulating Hormone (FSH)*

Elevated serum FSH levels are usually found in patients with non-obstructive azoospermia; however, a normal or near-normal serum FSH concentration does not always guarantee normal spermatogenesis [37]. Indeed, patients with NOA due to early maturation arrest may have low normal serum FSH level, despite having the worst chance of sperm retrieval [24]. The poor predictive ability of FSH on the chance of sperm retrieval was shown by a study on 792 men undergoing mTESE, which provided the counterintuitive demonstration that higher FSH levels were associated with greater chances of SSR [38]; a neural computational model built on 1026 men with NOA confirmed that relying on serum FSH level to counsel patients with NOA is no more accurate than flipping a coin [8]. Other studies challenged these results, but their sample size was not large enough to detect a true association.

Since FSH level correlates with the number of spermatogonia and, to a lesser extent, primary spermatocytes [39], relying on its serum levels to counsel patients with NOA about their probability of SSR may be misleading: patients with MA and HS, as well as those with SCO with or without foci of hypospermatogenesis (focal SCO), may have comparable FSH levels, but their probability of SSR differs significantly. Indeed, a prediction model built on a development ( *N* = 558) and a validation set ( *N* = 695) of patients with NOA demonstrated that serum FSH level is unable to predict histopathological subcategories such as MA and focal SCO, and has low sensitivity (40.9%) and specificity (46.8%) in predicting HS and SCO, respectively [40]. These data reinforce older data against the use of basal FSH level as a predictor of SSR in patients with NOA, and should discourage further evaluation of serum FSH as marker of residual spermatogenesis in these patients.
